System and method for generating images of subsurface structures

1. A system for generating images related to a subsurface region of interest, comprising:

a data storage device having seismic data and an earth model related to the subsurface region of interest;

at least one central processing unit (CPU) in communication with the data storage device, the CPU comprising at least one processing core and associated CPU memory;

at least one external co-processor in communication with the CPU, the external co-processor comprising at least one external co-processor processing core directly coupled to a corresponding local memory, and wherein the external co-processor processing core comprises computer programmable code for forward propagating a source wavefield using the earth model at a first time interval; transferring, at a second time interval, the forward propagated source wavefield to the CPU for compression and external storage; backward propagating the seismic data at the first time interval to derive backward propagated receiver wavefield; and transferring, at the second time interval, the backward propagated receiver wavefield to the CPU;

wherein the CPU processing core is programmed to retrieve and decompress the forward propagated source wavefield data and apply imaging conditions to the decompressed forward propagated source wavefield and the backward propagated receiver wavefield to construct image data representative of the subsurface region of interest.

2. The system of claim 1 , further comprising a display device in communication with the CPU for displaying the image data representative of the subsurface region of interest.

3. The system of claim 1 , further comprising finite difference modeling code for forward and backward propagating of the source and receiver wavefield data.

4. The system of claim 1 , further comprising asynchronous communication code for efficient exchange of data between the CPU and the external co-processor.

5. The system of claim 4 , wherein the asynchronous communication code comprises ghostpoints.

6. The system of claim 4 , wherein the asynchronous communication code comprises wavefield snapshots.

7. The system of claim 1 , further comprising 2-pass input/output bound code for implementing a Reverse Time Migration algorithm, realized through asynchronous data input/output and imaging concurrent with computation on co-processors.

8. A computer-implemented method of generating images related to a subsurface region of interest, comprising:

accessing, via a central processing unit (CPU), seismic data and an earth model related to the subsurface region of interest;

forward propagating a source wavefield using the earth model at a first time interval via at least one external co-processor coupled to the CPU;

transferring, at a second time interval, the forward propagated source wavefield to the CPU for compression and external storage;

backward propagating the seismic data at the first time interval via the external co-processor to derive backward propagated receiver wavefield;

transferring, at the second time interval, the backward propagated receiver wavefield to the CPU;

retrieving the stored forward propagated source wavefield;

decompressing the retrieved forward propagated source wavefield via the CPU; and

via the CPU, applying imaging conditions to the decompressed forward propagated source wavefield and backward propagated receiver wavefield to construct image data representative of the subsurface region of interest.

9. The method of claim 8 , further comprising displaying the image data representative of the subsurface region of interest.

10. The method of claim 8 , wherein the forward and backward propagating steps comprise finite difference modeling of the source and receiver wavefield.

11. The method of claim 8 , further comprising asynchronously communicating between the CPU and the external co-processor for efficient exchange of data.

12. The method of claim 11 , wherein the step of asynchronously communicating between the CPU and the external co-processor comprises using ghostpoints.

13. The method of claim 11 , wherein the step of asynchronously communicating between the CPU and the external co-processor comprises using wavefield snapshots.

14. The method of claim 11 , further comprising using a 2-pass forward propagation of the source wavefield followed backward propagating the seismic data as part of a Reverse Time Migration algorithm, realized through asynchronous data input/output and imaging concurrent with computation on co-processors.

15. The method of claim 8 , further comprising iterating full waveform inversion to utilize gradient information to update subsurface medium properties.

16. A method for propagating seismic wavefields, comprising:

initiating, via a central processing unit (CPU), a plurality of computational blocks in at least one external co-processor coupled to the CPU for coordinating computational functions related to the wavefield in the external co-processor;

initiating, via a central processing unit (CPU), a plurality of data transfer blocks in the external co-processor to coordinate data transfer functions related to the wavefield between the CPU and external co-processor;

controlling the propagating of the wavefield via a central processing unit (CPU) via a CPU subthread; and

concurrently performing, asynchronously, the computational and data transfer functions via corresponding streams without interfering with the CPU subthread.

17. The method of claim 16 , further comprising the step of combining at least two numerical operators in a single computational kernel to be executed by the external co-processor.

18. The method of claim 16 , wherein the step of performing the computational and data transfer functions comprises finite difference modeling in respect to wave propagation.